Regeneration

are used for procuring food, and as organs of offence and defence; the second and third pairs are used for walking. The following two pairs, that correspond to the last two pairs of walking legs of crabs and crayfishes, are small, and are used by the animal in bracing itself against the shell. The first three pairs of legs have an arrangement at the base, the “breaking-joint,” by means of which the leg is thrown off, if injured. The last two pairs of thoracic legs cannot be thrown off. The first three pairs of legs are often lost under natural conditions. In an examination of 188 individuals I found that 21 (or 11 per cent) had lost one or more legs. If one of the first three legs is injured, except in the outer segment, it is thrown off at the breaking-joint, and a new leg regenerates from the broken-off end of the stump that is left. The new leg does not become full size, and is of little use until the crab has moulted at least once. The leg breaks off so close to the body, and the part inside of the breaking-joint is so well protected by the bases of the other legs, that it is scarcely possible that the leg could be torn off inside of the breaking-joint, and, as a matter of observation, all crabs that are found regenerating their legs under natural conditions do so from the breaking-joint. If, however, by means of small scissors, the leg is cut off quite near the body, a new leg regenerates from the cut-end, even when the leg is cut off at its very base. The breaking-joint would thoroughly protect from injury the part of the leg that lies nearer to the body, and yet from this inner part a new leg is regenerated. Moreover, the new leg is perfect in every respect, even to the formation of a new breaking-joint. In this case we have a demonstration that there need be no connection between the liability of a part to injury and its power of regeneration.

In still another way the same thing may be shown. If the crab is anæsthetized, and a leg cut off outside of the breaking-joint, it is not, at the time, thrown off—the nervous system, through whose action the breaking off takes place, being temporarily thrown out of order. After recovery, although the leg is thrown off in a large number of cases, it is sometimes retained. In such cases it is found that from the cut-end the missing part is regenerated. In this case also we find that regeneration takes place from a part of the leg that can never regenerate under natural circumstances.

The third and fourth legs of the hermit-crab cannot be thrown off, but they have the power of regeneration at any level at which they may be cut off. They are in a position where they can seldom be injured, and I have never found them absent or injured in crabs caught in their natural environment. The soft abdomen is protected by the snail’s shell. At the end of the abdomen the last pair of abdominal appendages serve as anchors to hold the crab in the shell. These appendages are large and very hard, and can seldom be injured unless the abdomen itself is broken, and under these circumstances the crab dies. Yet if these appendages are cut off they regenerate perfectly, and after a single moult cannot be distinguished from normal ones.

The more anterior abdominal appendages are present only on one side of the adult, although they are present on both sides of the larva, and, to judge from a comparison with other crustacea, these appendages have degenerated completely on one side, and have become rudimentary in the male, even on the side on which they are present. They too will regenerate if they are cut off. In the female these appendages are used to carry the eggs, and are, therefore, of use. They also have a similar power of regeneration. The maxillæ and maxillipeds of the hermit-crab have likewise the power of regeneration, as have also the two pairs of antennæ and the eyes.

In other decapod crustacea also it has been shown that the power of regeneration of the appendages is well developed. It has been long known that the crayfish and the lobster can regenerate lost parts. The first pair of legs, or chelæ, in these forms has a breaking-joint, at which the leg can be thrown off, yet in the crayfish I have seen that if the leg is cut off inside of the breaking-joint it will regenerate. The four pairs of walking legs do not possess a breaking-joint, but may be thrown off in some cases at a corresponding level. They regenerate from this level, as well as nearer the body and farther beyond this region. Przibram has recently shown that, in a number of crustacea, regeneration of the appendages takes place, even when the entire leg is extirpated as completely as possible.

Newport has shown that the myriapods can regenerate their legs, and it is known that several forms have the power of breaking off their legs in a definite region at the base if the legs are injured, and I have observed in Cermatia forceps that this takes place even when the animal is thrown into a killing fluid. Newport (’44) has also shown that when the legs of a caterpillar are cut off new ones regenerate during the pupa stage. It has been long known[42] that the legs of mantis can regenerate, and Bordage, who has recently examined the question more fully, has shown that a breaking-joint is present at the base of the leg. The tarsus of the cockroach also regenerates, producing only four, instead of the five, characteristic segments.[43]

A number of writers have recorded the regeneration of the legs of spiders.[44] Schultz, who has recently examined more thoroughly the regeneration of the legs in some spiders, finds that the leg is renewed if cut off at any level. He removed the leg most often at the metatarsus, but also at the tibia, and generally between two joints. In some cases the leg was cut off at the coxa, at which level it is generally found to be lost under natural conditions. Wagner observed in tarantula that when the leg is removed at any other place than at the coxa, the animal brings the wounded leg to its jaws, and bites it off down to the coxa. In the Epeiridæ, that Schultz chiefly made use of, this never happened. He observed, however, even in these forms, that when the leg is cut off at the coxa it regenerates better than

when cut off at any other level. Schultz states that we see here an excellent example of how regeneration is influenced by natural selection, since regeneration takes place best where the leg is most often broken off. On the other hand, the author hastens to add that since regeneration also takes place when the leg is cut off at any other level, this shows that the power to regenerate is characteristic of all parts of the organism, and is not merely a phenomenon of adaptation, as Weismann believes. It seems highly improbable that a spider could ever lose a leg in the middle of a segment, i.e. between two joints, since the segments are hard and strong and the joints much weaker; but nevertheless the leg has the power to regenerate also from the middle of the segment, if cut off in this region.

The formation of the new part takes place somewhat differently, according to Schultz, when the leg is amputated between two segments than when cut off at the coxa. In the latter case, there is produced from the cut-end of the last segment a solid rod which, as it grows longer, bends on itself several times. Joints appear in the rod, beginning at the base. The leg is set free at the next moult. If the leg is cut off nearer the distal end a smaller rod is formed, that extends straight forward, or may be thrown into a series of folds. It lies, however, inside of the last segment, since the surface exposed by the cut is quickly covered over by a chitinous covering. The piece is set free at the next moult.

Loeb has found that if the body of the pycnogonid, Phoxichilidium maxillare, is cut in two there regenerates from the posterior end of the anterior half a new body-like outgrowth.

Without attempting to describe the many cases in worms and mollusks in which there is no obvious connection between the power of the part to regenerate and its liability to injury, but where it is more difficult to show that it may not exist, let us pass to an examination of the regeneration of the starfish. It has been known since the time of Réaumur that starfish have the power of regenerating new arms if the old ones are lost. It has been stated that in certain starfishes an arm itself can produce a new starfish,—Haeckel (’78), P. and R. Sarasin (’88), von Martens (’84), and Sars (’75),—but this has been denied by other observers. In several species of starfishes, the separated arm does not regenerate; but if a portion, even a small piece, of the disk is left with the arm, a new disk and arms may develop (Fig. 38, F). When the arm of Asterias vulgaris is injured it pinches off in many cases at its base, and a new arm grows out from the short stump that remains. When these starfishes regenerate new arms in their natural environment, the new arms almost always arise from this breaking region.[45] Thus King found out of 1914 individuals of Asterias vulgaris collected at random, 206, or 10.7 per cent, had one or more new arms, and all these except one arose from near the disk. In other species it appears that the outer portions of the arm may be broken off without the rest of the arm being thrown off. King has found that in asterias, regeneration takes place more rapidly from the base than at a more distal level. It may appear, at first thought, that the more rapid regeneration of the arm at the place at which it is usually thrown off may be associated with its more frequent loss at this region—in other words, that the more rapid regeneration has been acquired by the region at which the arm is generally broken off. This interpretation is, however, excluded by the fact that, in general, the nearer to the base the arm is cut off, so much the more rapid is its regeneration. In other words, the more rapid regeneration of the arm at the base is only a part of a general law that holds throughout the arm. If the proposition is reversed, and it is claimed that the arm has acquired the property of breaking off at the base, because it regenerates more rapidly at that level, the following fact recorded by King is of importance, viz. that, although the arm regenerates faster at the base, yet a new arm is not any sooner produced in this way, since there is more to be produced and the new arm from the base may never catch up to one growing less rapidly from a more distal cut-surface, but having a nearer goal to reach.

The results of our examination show that those forms that are liable to have certain parts of their bodies injured are able to regenerate not only these parts, but at the same time other parts of the body that are not subject to injury. The most remarkable instance of this sort is found in those animals having breaking-joints. In these forms, we find that regeneration takes place both proximal and distal to this region. If the power of regeneration is connected with the liability of a part to injury, this fact is inexplicable.

Turning now to the question as to whether regeneration takes place in those species that are subject to injury more frequently or better than in other species, we find that the data are not very complete or satisfactory for such an examination. It is not easy to ascertain to what extent different animals are exposed to injury. If we pass in review the main groups of the animal kingdom, we can at least glean some interesting facts in this connection.

In the protozoa nucleated pieces have been found to regenerate in all forms that have been examined, including amœba, difflugia, thalassicolla, paramœcium, stentor, and a number of other ciliate infusoria.

In the sponges it has been found by Oscar Schmidt that pieces may produce new individuals, but how widely this occurs in the group is not known. In the cœlenterates many forms are known to regenerate, and it is not improbable that in one way or another the process occurs throughout the group. The hydroid forms, hydra, tubularia, parypha, eudendrium, antennularia, hydractinia, podocoryne, etc., the jelly-fish, gonionemus, and certain members of the family Thaumantidæ, have been found to regenerate. Amongst the Scyphozoa, metridium, cerianthus, and the scyphistoma of aurelia regenerate, and the jelly-fishes belonging to this group have a limited amount of regenerative power.

In the platodes we find that all the triclads, thus far examined, including planaria, phagocata, dendrocœlum, and the land triclad, bipalium, regenerate. It has been shown that the marine triclads also regenerate, but less rapidly and extensively, while the marine polyclads have very limited powers of regeneration. The regeneration of the trematodes and cestodes has not, so far as I know, been studied, neither have the nematodes been examined from this point of view.

Some of the nemerteans regenerate, others do not seem to have this power. A small fresh-water form, tetrastemma, that I examined, did not regenerate, although some of the pieces, that were filled with eggs, remained alive for several months.

In the annelids we find a great many forms that regenerate—many marine polychæta have this power; all oligochæta that have been studied regenerate; both land forms, like lumbricus, allolobophora, etc., and fresh-water forms, like lumbriculus, nais, tubifex, etc.

In the crustacea the appendages have the power to regenerate in all the forms that have been examined.

Several kinds of myriapods, as well as a number of spiders, are known to regenerate their legs. In the insects, however, only a few forms are known to have this power,—caterpillars, mantis, and the cockroach. The large majority of insects, in the imago state, do not seem to be able to regenerate, although in a few cases regeneration has been found to occur.[46]

In the mollusks, regeneration of the head takes place under certain conditions. Spallanzani thought that if the entire head is cut off a new one regenerates. This conclusion was denied by at least eleven of his contemporaries, and confirmed by about ten others. It was found later that the result depends in part on the time of year and in part on the kind of snail. Carrière, who more recently examined the question, found that even under the most favorable conditions regeneration does not take place if the circumœsophageal nerve-commissure is completely removed with the head, but if a part remains, a new head develops. It has been stated that a new foot regenerates in helicarion, and I have found that the foot regenerates also in the fresh-water snails, physa, limnæa, and planorbis. If the margin of the shell of a lamellibranch or of a snail is broken off, it is renewed by the mantle. The arms of some of the cephalopods are known to regenerate, particularly the hectocotylized arm.

In all the main groups of echinoderms, with one possible exception, regeneration has been found to take place. Probably all starfishes and brittle-stars regenerate their arms, and even if cut in two or more pieces, new starfishes develop. The crinoids regenerate lost arms, and even parts of the disk; also the visceral mass. The holothurians have very remarkable powers of regeneration. In some forms regeneration takes place if the animals are cut in two, or even in more than two pieces. The remarkable phenomenon of evisceration that take place in certain holothurians, if they are roughly handled, or kept under unfavorable conditions, are well known and have been described by a number of writers. It has even been suggested that the holothurian may save itself by offering up its viscera to its assailant! Unfortunately for this view, it has been found that the viscera are unpalatable, at least to sea-anemones and to fishes. Ludwig and Minchin suggest that the throwing off of the Cuvierian organs, which are attached to the cloaca, is a defensive act, and if carried too far, according to the latter writer, the viscera may also be lost. The holothurians have remarkable recuperative powers and may regenerate new viscera in a very short time. The sea-urchins form, perhaps, an exception in this group, since there are no records of their regenerative power, but no doubt this is because they have not been as fully investigated as have other forms.

In the vertebrates the lower forms, amphioxus, petromyzon, and sharks, have not been studied in regard to their regenerative power. In the teleostean fishes the fins of a number of forms are known to regenerate. It is probable that this takes place in most members of the group.

In the amphibia we find a large number of forms that regenerate their limbs and tail, and other parts of the body, but limitations appear in certain forms. The rapid regeneration of the legs in the smaller urodeles has been often described. In larger forms it takes place more slowly, at least in large forms having large legs. In proteus the regeneration may extend over a year and a half, and in necturus it takes more than a year to make a new limb, at least in animals in confinement. In the large form, amphiuma, that has extremely small legs, regeneration takes place much more rapidly than in a form like necturus having much larger legs (Fig. 39).

In amphiuma the feet are not used by the animal as organs of locomotion, since they are too small and weak to support the heavy body. They can be moved by the animal in the same way that the feet are moved in other forms, and yet are useless for progression. It is said by Schreiber that the regeneration of the legs of Triton marmoratus is relatively very slight as compared with that of other forms. Fraisse also found in this form that an amputated leg did not grow again, only a deformed stump being produced. The tail also is said to regenerate to only a slight extent, but, so far as I know, there is nothing peculiar in the life of this form that makes it less liable to injury than other large urodeles.[47] Weismann cites the case of proteus, which is said also to regenerate less well than do other forms. It lives in the caves of Carniola, where there are few other animals that could attack or injure it, and to this immunity is ascribed its lack of power of regeneration; yet Goette states that he observed a regenerating leg in this form, but that the process was not complete after a year and a half. In necturus also, which is not protected in any way, regeneration is equally slow. Frogs are unable to regenerate their limbs, although they are sometimes lost, but the larval tadpole can regenerate at least its hind legs. In the lizards the tail regenerates, but at present we do not know of any connection between this condition and the liability of certain forms to injury. Turtles and snakes do not regenerate their tails. I do not know of any observations on crocodiles.

In birds, the legs and wings are not supposed to have the power to regenerate,[48] but in two forms[49] at least the beak has been found to possess remarkable powers of regeneration. There are a few very dubious observations in regard to the regeneration in man of superfluous digits that had been cut off.[50]

These examples might be added to by others in the groups cited, and also by examples taken from the smaller groups of the animal kingdom, but those given will suffice, I think, to show that the power to regenerate is characteristic of entire groups rather than individual species. When exceptions occur, we do not find them to be forms that are obviously protected, but the lack of regeneration can rather be accounted for by some peculiarity in the structure of the animal. If this is borne in mind, as well as the fact that protected and unprotected parts of the same animal regenerate equally well, there is established, I think, a strong case in favor of the view that there is no necessary connection between regeneration and liability to injury. We may therefore leave this side of the question and turn our attention to another consideration.

It will be granted without argument that the power of replacement of lost parts is of use to the animal that possesses it, especially if the animal is liable to injury. Cases of usefulness of this sort are generally spoken of as adaptations. The most remarkable fact in connection with these adaptive responses is that they take place, in some cases at least, in parts of the body where they can never, or at most very rarely, have taken place before, and the regeneration is as perfect as when parts liable to injury regenerate. Another important fact is that in some forms the regeneration is so slow that if the competition amongst the animals was very keen those with missing legs, or eyes, or tails, would certainly succumb; yet, if protected, they do not fail to regenerate. If, therefore, the animal can exist through the long interval that must elapse before the lost part regenerates, we cannot assume that the presence of the part is of vital importance to the animal, and hence its power to regenerate could scarcely be described as the result of a “battle for existence,” and without this principle “natural selection” is powerless to bring about its supposed result.

It is extremely important to observe that some cases, at least, of regeneration are not adaptive. This is shown in the case where a new head regenerates at the posterior end of the old one in Planaria lugubris, or where a tail develops at the anterior end of a posterior piece of an earthworm, or when an antenna develops in place of an eye in several crustacea. If we admit that these results are due to some inner laws of the organism, and have nothing to do with the relation of the organism to its surroundings, may we not apply the same principle to other cases of regeneration in which the result is useful?

So firm a hold has the Darwinian doctrine of utility over the thoughts of those who have been trained in this school, that whenever it can be shown that a structure or a function is useful to an animal, it is without further question set down as the result of the death struggle for existence. A number of writers, being satisfied that the process of regeneration is useful to the animal, have forthwith supposed that, therefore, it must have been acquired by natural selection. Weismann has been cited as an example, but he is by no means alone in maintaining this attitude. It would be entirely out of place to enter here into a discussion of the Darwinian theory, but it may be well worth while to consider it in connection with the problem of regeneration.

We might consider the problem in each species that we find capable of regenerating; or, if we find this too narrow a field for our imagination, we might consider the process of regeneration to have been “acquired by selection in the lower and simpler forms,” and trace its subsequent progress as it decreased in the course of phylogeny “in correspondence with the increase in complexity of organization,” or with the decrease of exposure to injury. At the risk of adopting the narrower point of view I shall confine the discussion to the possibility of regeneration being acquired, or even augmented, through a process of natural selection in any particular species.

The opportunity to regenerate can only occur if a part is removed by accident or otherwise. On the Darwinian theory we must suppose that of all the individuals of each generation that are injured, in exactly the same part of the body, only those have survived or have left more offspring that have regenerated. In order that selection may take place, it must be supposed that amongst these individuals injured in exactly the same region, regeneration has been better in some forms than in others, and that this difference is, or may be, decisive in the competition of the forms with each other. The theory does not inquire into the origin of this difference between individuals, but rests on the assumption of individual differences in the power to regenerate, and assumes that these differences can be heaped up by the survival and inbreeding of the successful individuals; i.e. it is assumed that, by this picking out or selection through competition in each generation of the individuals that regenerate best, the process will become more and more perfectly carried out in the descendants, until at last each part has acquired the power of complete regeneration.

There are so many assumptions in this argument, and so many possibilities that must be realized in order that the result shall follow, that, even if the assumptions were correct, one might still remain sceptical in regard to the possibilities ever becoming realized. If we examine somewhat more in detail the conditions necessary to bring about this supposed process, we shall find ample grounds for doubt, and even, I think, for denial that the results could ever have been brought about in this way.

In the first place, the assumption that the regeneration of an organ can be accounted for as a result of the selection of those individual variations that are somewhat more perfect, rests on the ground that such variations occur, for the injury itself that acts as a stimulus is not supposed to have any direct influence on the result, i.e. for better or worse. All that natural selection pretends to do is to build up the complete power of regeneration by selecting the most successful results in the right direction. In the end this really goes back to the assumption that the tissue in itself has power to regenerate more completely in some individuals than in others. It is just this difference, if it could be shown to exist, that is the scientific problem. But, even leaving this criticism to one side, since it is very generally admitted, it will be clear that in many cases most of the less complete stages of regeneration that are assumed to occur in the phyletic series could be, in each case, of very little use to the individual. It is only the completed organ that can be used; hence the very basis of the argument falls to the ground. The building up of the complete regeneration by slowly acquired steps, that cannot be decisive in the battle for existence, is not a process that can be explained by the theory.

There is another consideration that is equally important. It is assumed that those individuals that regenerate better than those that do not, survive, or at least have more descendants; but it should not be overlooked that the individuals that are not injured (and they will belong to both of the above classes) are in even a better position than are those that have been injured and have only incompletely regenerated. The uninjured forms, even if they did not crowd out the regenerating ones, which they should do on the hypothesis, would still intercross with them, and in so doing bring back to the average the ability of the organism to regenerate. Here we touch upon a fatal objection to the theory of natural selection that Darwin himself came to recognize in the later editions of the Origin of Species, namely, that unless a considerable number of individuals in each generation show the same variation, the result will be lost by the swamping effects of intercrossing. If this be granted, there is left very little for selection to do except to weed out a few unsuccessful competitors, and if the same causes that gave origin to the new variation on a large scale should continue to act, it will by itself bring about the result, and it seems hardly necessary to call in another and questionable hypothesis.

Finally, a further objection may be stated that in itself is fatal to the theory. We find the process of regeneration taking place not only at a few vulnerable points, but in a vast number of regions, and in each case regenerating only the missing part. The leg of a salamander can regenerate from every level at which it may be cut off. The leg of a crab also regenerates at a large number of different levels, and apparently this holds for all the different appendages. If this result had been acquired through the action of natural selection, what a vast process of selection must have taken place in each species! Moreover, since the regeneration may be complete at each level and in each appendage without regard to whether one region is more liable to injury than is another, we find in the actual facts themselves nothing to suggest or support such a point of view.

If, leaving the adult organism, we examine the facts in regard to regeneration of the embryo, we find again insurmountable objections to the view that the process of regeneration can have been produced by natural selection. The development of whole embryos from each of the first two or first four blastomeres can scarcely be accounted for by a process of natural selection, and this is particularly evident in those cases in which the two blastomeres can only be separated by a difficult operation and by quite artificial means. If a whole embryo can develop from an isolated blastomere, or from a part of an embryo without the process having been acquired by natural selection, why apply the latter interpretation to the completing of the adult organism?

Several writers on the subject of regeneration in connection with the process of autotomy (or the reflex throwing off of certain parts of the body) have, it seems to me, needlessly mixed up the question of the origin of this mechanism with the power of regeneration. If it should prove true that in most cases the part is thrown off at the region at which regeneration takes place to best advantage, it does not follow at all that regeneration takes place here better than elsewhere, because in this region a process of selection has most often occurred. The phenomenon of regeneration in the arm of the starfish, that has been described on a previous page, shows how futile is an argument of this sort. If, on the other hand, the autotomy is supposed to have been acquired in that part of the body where regeneration takes place to best advantage, then our problem is not concerned with the process of regeneration at all, but with the origin of autotomy. If the attempt is made to explain this result also as the outcome of the process of natural selection acting on individual variations, many of the criticisms advanced in the preceding pages against the supposed action of this theory in the case of regeneration can also readily be applied to the case of autotomy. In Chapter VIII, in which the theories of autotomy are dealt with, this problem will be more fully discussed.

CHAPTER VI

REGENERATION OF INTERNAL ORGANS. HYPERTROPHY. ATROPHY

The experiments made by Ponfick (’90) on the regeneration of the liver in dogs and in rabbits gave the most striking results. Ponfick found after removal of a fourth, or of a half, or even, in a few successful operations, of three-fourths of the liver, that, in the course of four or five weeks, the volume of the remaining part increased, and in the most extreme case, to three times that of the piece that had been left in the body. The first changes were found to have begun as early as thirty hours after the operation, when the liver cells had begun to divide. The maximum number of dividing cells was found about the seventh day, and then decreased from the twentieth to the twenty-fifth day, but cells were found dividing even on the thirtieth day. These dividing cells appeared everywhere throughout the liver, and were no more abundant at the cut-edges than elsewhere. There takes place, in consequence, an increase in the volume of the liver, rather than a replacement of the part that is removed. The increase takes place in the cells of the old part, the lobules swelling up to two, three, or even four times their former size. No new liver lobules seem be formed. The old tubules of the liver also become larger, owing to an increase in the number of their cells. Since the change takes place in the old part, and is due to an increase in size of the lobules, tubes, etc., the process is spoken of as one of hypertrophy rather than of regeneration.

Kretz found a case in which the entire parenchyma of the liver seemed to have been destroyed, presumably by a poison from some micro-organism, and later a regeneration of the tissue had taken place. If this conclusion is correct, it shows that sometimes an internal organ may meet with an injury that does not directly destroy the rest of the body, and the animal may survive.

The regeneration of the salivary gland of the rabbit described by Ribbert is another example of an internal organ that can seldom be injured, and yet can be replaced after artificial removal. Weismann (’93) has recorded an experiment in which half of a lung of triton was cut off. After fourteen months the lung had not been restored in four individuals, and in one “it was doubtful whether a growth of the lung had not taken place, but even in this case it had not recovered its long, pointed form.”

The regeneration of the eye in triton was first made known by Bonnet. The right eye was partly cut out, and after two months it had completely regenerated. Blumenbach, in 1784, removed the anterior part of the bulb of the eye of “Lacerta lacustris.” Six months later a smaller bulb was present. Phillipeaux (’80) found that if the eye of an aquatic salamander was not entirely removed, a new eye regenerated; but if the eye was completely extirpated a new eye did not appear. Colucci, in 1885, described the regeneration of the lens of the eye of triton from the edge of the optic cup. Wolff, later, independently, discovered the same fact, and it has been more recently confirmed by E. Müller (’96), W. Kochs (’97), P. Rothig (’98), and Alfred Fischel (’98). The most important part of this discovery is that the new lens develops from the margin of the optic cup, and not from the outer ectoderm, as it does in the embryo. This result will be more fully discussed in a later chapter. It is highly probable in this case that the regeneration stands in no connection whatsoever with the liability of the eye to injury, for of the large number of salamanders that have been examined, none has been found with the eye mutilated. The position of the eye is such that it is well protected from external injury, and the tough cornea covering its outer surface would also further protect it from accidental injury. When we recall the high degree of structural complexity of the eye, its capacity to regenerate, if only a portion of the bulb is left, and its power to replace the lens if this is removed are certainly very remarkable facts. We find here, I think, an excellent refutation of the incorrectness of the general assumption of a connection between regeneration and liability to injury. Moreover, since there is no evidence whatsoever to show that the eyes in these animals are ever subject to diseases caused by bacteria, and much evidence to show that they are not so injured, we are still further confirmed in our general conclusion.

It has been known for a long time that even in man the lens of the eye is sometimes regenerated after its removal. The regeneration has been supposed to take place from the old capsule of the lens, or possibly from a piece of the lens left after the operation; but whatever its origin, the fact of its regeneration in man, and in other mammals also, is a point of some interest in this connection.

Podwyssozki (’86) found that regeneration may take place in the kidney of certain mammals,—best in the rat, more slowly in the rabbit. The restoration of the lost part takes place first by replacement of the epithelium. The old canals may then push out into the connective tissue that accumulates in the new part, but there is no new formation of canals or of glomeruli. According to Podwyssozki the regeneration of the kidney is less complete than that of any other gland. Peipers has reinvestigated the subject, and his results agree in the main with those just given. He finds in addition that new canals may grow out from the old ones into the new part.

Podwyssozki and Ribbert (’97) have found that the salivary gland has a remarkable power of regeneration. Ribbert removed a half (or even more than this) of the salivary gland of the rabbit. In the course of two or three weeks new material had developed over the cut-surface. In one case at least five-sixths of the gland had been taken out, and at the end of three weeks the gland had regenerated to its full size. Microscopic examination showed that the greater part of the gland was made up of new lobes, some of which were as large as, others smaller than, the normal lobes. The new part contained new tubes with terminal acini. These had arisen from the tubes of the old part. The connective tissue of the new part also came from that of the old. In this case a true process of regeneration takes place from the cut-surface; in addition a certain amount of enlargement, or hypertrophy, also takes place in the old part. Ribbert believes there is a connection between the process of hypertrophy and of regeneration of such a kind that the more active the one, the less active the other.

Regenerative changes are known to occur in other internal organs besides these glandular ones. Broken bones are united, if brought in contact, by a process that involves a certain amount of regeneration. Although new bony tissue may be formed at the region of union, the bones of mammals and of birds do not seem able to complete themselves, if a part is removed, except to a limited extent. While the broken bones of the leg or of the arm have the power of reuniting if held for some time in place, yet in nature this condition can seldom be fulfilled, and the animal with a broken leg or wing will most probably be killed. Nevertheless, since the bones have this power at whatever level they may be broken (but only if they are kept together artificially), the process can scarcely have been acquired through the liability of the parts to injury. We find here another instance of a useful process existing in animals, but one that could not have been acquired by exposure of the part to injury. It is probable that this same property is found in all the bones of the body,—in those that may occasionally be injured, and in those that are not.

The muscles have also the power of regenerating, although few experiments have been made except in those forms in which the whole leg can regenerate, yet there are a few observations that show that even in mammals, in which the leg or the arm cannot regenerate as a whole, a certain amount of regeneration of the muscles themselves may take place.

It has been known for a long time that if a nerve is cut a new nerve grows out from the cut-end, and may extend to the organs supplied by that nerve. The process takes place more successfully if the peripheral part is left near the cut-end from which the new nerve grows. Whether this old part only serves to guide the new part to its proper destination, or whether it may also contribute something to the new nerve, as, for instance, cells for the new sheath, is not finally settled. The general opinion in regard to the origin of the new nerve fibres is that the central axis or fibril grows from the cut-end. That this power could have been acquired for each nerve as a result of its liability to injury is too improbable to discuss seriously.

The central nervous system of the higher vertebrates seems to have very little power of regeneration, and although in some cases a wounded surface may be covered over and a small amount of connective tissue be formed, the development of new ganglion cells does not seem to occur. In other animals, as the earthworm, planarian, and even in the ascidian, as shown by Loeb, a new entire brain may develop after the removal of the old brain, or of that part of the body in which it is contained.

This examination of the power of regeneration of internal organs in the vertebrates has shown that it is highly improbable that there can be any connection between their power of regeneration and their liability to injury. That the internal organs may be occasionally injured by bacteria, or by poisons made in the body, may be admitted, but that injuries from this source have been of sufficient frequency to establish a connection, if such were indeed possible, between their power of regeneration and their liability to injury from these causes is too improbable a view to give rise to much doubt. These results taken in connection with those discussed in the preceding chapter go far toward disproving the view that the power of regeneration has a connection with the liability of a part to injury.

HYPERTROPHY

The hypertrophy, or unusual enlargement, of organs has long attracted the attention of physiologists, and the extremely interesting observations and experiments that have been made in this connection have an important although an indirect bearing on the problem of regeneration. Ribbert, as has been pointed out, holds that the processes of hypertrophy and of regeneration stand in a sort of inverse relation to each other, but it is doubtful, I think, if any such general relation exists. Two kinds of hypertrophy are now generally distinguished: functional hypertrophy, which takes place when a part becomes enlarged through use; and compensating hypertrophy, which takes place when one organ being removed another enlarges. The enlargement in the latter case may, of course, be brought about by the increased use of the parts that enlarge, but as this is not necessarily the case, the distinction between the two processes is a useful one. The causes of compensating hypertrophy are by no means simple, and several possibilities have been suggested to account for the enlargement. The best ascertained facts in connection with hypertrophy relate almost entirely to man and to a few other mammals.[51]

By hypertrophy is meant an increase of the substance of which an organ is composed. Swelling due to the imbibition of water or of blood-serum is not, in a technical sense, a process of hypertrophy. Virchow distinguishes two kinds of hypertrophy: (1) Hypertrophy in a narrower sense in which the enlargement is due to an increase in the size of the cells of which an organ is composed. This enlargement of the individual cells leads of course to an increase in the size of the whole organ. (2) Hyperplasy due to an increase in the number of cells of which an organ is composed, which also causes an enlargement of the whole organ if the cells retain the normal size. The division into functional and compensating hypertrophy given above is a physiological distinction, and both of these processes might occur in Virchow’s subdivisions.

Giants may be looked upon as hypertrophied individuals, since all the organs of the body are larger than the normal. The enlargement is, in this case, not due to external influences, but to some peculiarity of the organism itself. Whether the size is due to more cells being present, as seems probable, or to the cells being larger, or to both, has not, so far as I know, been determined for man. In a mollusk, Crepidula fornicata, in which large and small adult individuals occur, it has been shown by Conklin (’98) that the difference is due entirely to the larger number of cells in the larger individual. In this case external conditions, in so far as they retard the maximum possible growth of the individual, are responsible for the differences in size. The distinction is, in this case, rather between large normal individuals and dwarfs, than between giants and normal or average individuals.

The voluntary muscles of the body of man grow larger, and may be said to hypertrophy, as a result of doing certain kinds of work. The muscles of the hand and arm grow large through use, and become smaller again if not used; but the muscles of the fingers of a musician do not hypertrophy, although the total amount of work done may be very large. It is only when muscular work is done against great resistance that enlargement of the muscles takes place. The factors that may bring about the enlargement will be discussed later.

The kidneys seem to give the most satisfactory evidence of compensating hypertrophy. Nothnagel[52] states that it has been shown in man, in the rabbit, and in the dog, that when one kidney has been removed the other enlarges; and that this takes place both for young animals, in which the kidneys have not reached their full size, and in adult animals, in which the remaining kidney becomes larger than normal. In the adult the enlargement is due to hypertrophy, in Virchow’s sense, in the tubules and in the epithelium of the canals. In the young animal there is, in addition, a hyperplastic growth that leads to an increase in the number of glomeruli, etc.

Experiments have shown that the same amount of urea is excreted by the animal after the removal of one kidney as before; in fact, this is true immediately after the operation, before any increase in the size of the organ has taken place. This means that, under normal conditions, the kidneys do not perform their maximum of work. It is important to observe in this connection that the remaining kidney gets more blood than it would get if the other were present. Nothnagel sums up the changes that take place in this way: First, the removal of one kidney; second, an increase in the flow of blood in the remaining kidney; third, an increase in the functional activity and excretion of this kidney; fourth, along with the increase in the flow of blood, there is a necessary increase in the amount of food that is brought to the kidney in the blood; fifth, this food is taken up in larger amount than before by the cells, which leads to an increase in the growth of the cells, which produces hypertrophy. The increase in size, looked at from this point of view, Nothnagel says, has nothing mysterious about it. The enlargement seems to be an adaptation; but the enlargement does not take place because it is an adaptive process, but because it cannot be helped under the conditions that arise. We shall return again to Nothnagel’s interpretation, when we come to consider other views.

Experiments of the sort just described are most easily carried out on the paired organs of the body, such as the salivary glands, the tear glands, the mammæ of the female, and the testes of the male. In regard to the latter two organs the evidence, especially in the case of the testes, is conflicting, but the recent experiments of Ribbert seem to give definite results. Nothnagel had found that after the removal of one testis there is no hypertrophy of the other. He pointed out that this result does not stand in contradiction to his hypothesis in regard to the kidneys, for the loss of one testis does not lead to a greater functional activity in the other. Each acts for itself alone. The result shows further, he adds, that the process of hypertrophy is not an adaptive one, but a physical or a physiological process. Ribbert on the contrary thinks that even Nothnagel’s statistics give evidence of hypertrophy, and Ribbert’s own experiments give unmistakable evidence of a considerable enlargement of the remaining testis. In his experiments, young rabbits were used that were born of the same mother and in the same litter. One of the testes was removed from some of the individuals, and after some months the remaining testis was taken out and its weight compared with that of the control animal. In sixteen out of seventeen experiments there was found to be a noticeable increase in the single testis as compared with either testis of the control animal. The results show that in some cases the single testis weighs almost as much as the two together of the control animal. It is important also to notice that in this case the enlargement has taken place in an organ that has not been active, as was the case with the kidney.

Ribbert has also shown that hypertrophy takes place in the mammæ of the rabbit after the removal of some of them. Five out of the eight mammæ were removed in three cases, and seven out of the eight in two other cases from young rabbits about two months old. Ribbert found that if the operator is not careful to remove completely all the tissue of a mamma an active regenerative process takes place from the part that remains. After five and a half months the single remaining mamma of one animal measured six and one-half by three and four-fifths centimetres, and the corresponding one in the control animal five and three-fourths by three and one-half centimetres. The glandular tissue was also found less developed in the control animal.

In another experiment the rabbit experimented upon bore young when it was six and a half months old. Soon after the birth of the young and before the mamma had been used the animal was killed and the single mamma that had been left was measured. It was much enlarged and projected more than the normal mammæ. It measured nine by five centimetres. In a normal control animal[53] the corresponding mamma measured seven by five centimetres. The number of acini was in the proportion of sixteen in the animal operated upon to ten in the normal. The results show a distinct compensating hypertrophy, due to a hyperplastic increase in the number of elements of the gland.

A further example of compensating hypertrophy has been found after the removal of the spleen, when the lymphatic glands of other parts of the body become enlarged. There are also observations which go to show that after the removal of some of the lymphatic glands others undergo an enlargement.

Ziegler[54] has given a critical review of the various opinions and hypotheses that have been advanced to account for the process of hypertrophy. According to Cohnheim[55] hypertrophy in bones, muscles, spleen, and glands is due to hyperæmia, i.e. increased blood supply. He thinks that neither mechanical nor chemical stimuli can cause directly new processes of growth. Recklinghausen[56] thinks that hypertrophy is not due to any extent to an increase in the food supply. Samuel[57] explains hypertrophy as due to a removal of, or to a decrease in, the resistance to growth and also to the influence of the nerves. Klebs[58] thinks that three factors enter into the problem, (a) inherited peculiarities, (b) overfeeding, (c) a removal of the controlling influences. Weigert believes that reparative processes are due to the removal of influences that prevent growth, and not to a direct stimulus. He thinks that a stimulus may start a functional act, but can never start a nutritive or a formative one. Good nourishment, for instance, may bring a tissue to a maximum development that is predetermined by innate peculiarities, but “idioplastic forces” are not thereby increased. Pekelharing[59] thinks that hypertrophy is due to a disappearance of a resistance to growth, and also to a stimulus causing proliferation.

We see from these various opinions how little is really known; how little has been determined as yet by experiment as to the causes that bring about hypertrophy. Many of the views are more or less plausible in the absence of direct, experimental evidence, but it remains for the future to decide as to the correctness of all of them. They are valuable as suggestions, in so far as they show the different possibilities that must be taken into account.

Ziegler first advocated the view, in the first edition of his Lehrbuch, that hypertrophy is due to a lessening of the resistance to growth. He thinks that while hyperæmia and transudation may support the new growth, they are never the only cause of the formation of new tissue. While Virchow’s view that any injury to the body or to an organ excites proliferation finds support in the work of Stricker and Grawitz, yet the view has been combated by Cohnheim and by Weigert, and is no longer held by many pathologists. Ziegler points out that as a result of his own work, and that of his students, traumatic and chemical lesions are not followed at once by new growth of the tissue, but by degeneration of the tissue, and by changes in the circulation that lead to exudations. The new growth begins, at the earliest, eight hours after the operation, and generally only after twenty-four hours. Also after mechanical, chemical, or thermal injuries, a long interval elapses before phenomena of growth begin. The injury itself does not appear to produce the growth, but brings about those conditions that lead to cell-multiplication. Ziegler discusses what is meant by the idea of a lessening of the resistance to growth. He himself does not mean by this that hypertrophy depends on changes in the physical conditions, because it is known that living phenomena are the outcome of chemical processes and it is, therefore, à priori probable that the effect is brought about by chemical substances in the fluids of the tissues. These substances affect functional actions, and may even bring about regenerative changes. This action of chemical substances on the formative activity of the cell is theoretically possible in either of two ways; first, chemical substances of definite concentration are set free, or, second, chemical substances are present in the normal condition that prevent proliferation, but if their influence should be counteracted by other substances the conditions become favorable to growth. It is known in the case of certain unicellular organisms, that derive their nourishment from the surrounding medium, that their increase in number may be retarded by the presence of certain chemical substances. It is also known that certain organisms may themselves produce chemical substances that prevent their own multiplication. It is, therefore, at least conceivable that after a part has been injured a new substance may be produced that acts upon and destroys in the organ itself the substances there present that have prevented its further growth. The other interpretation is that in the breaking down of the tissue of the organ a substance is produced that excites the cells to proliferation.

Klebs suggested that the accumulation of the leucocytes at the wounded surface may act as a stimulus to growth, and that the chromatin of their nuclei might be absorbed by the cells of the tissue, and combining with the nuclei of these cells bring about the new growth. But Ziegler points out that we now know that although the leucocytes are dissolved and absorbed over the wounded surface, no process of absorption, of the sort postulated by Klebs, takes place. Ziegler thinks that Nothnagel is wrong in supposing that an increase in the blood supply, bringing with it an increase in the nourishment, can account for the hypertrophy of the kidney. On the contrary he believes that the growth is the result of an increase in the function of the organ due to the increase of the chemical substance, urea, that is brought to the secreting cells. The muscles of the body also hypertrophy as a result of their activity and not as a result of the additional blood supply.

In connection with these problems of hypertrophy it may be pointed out that, under certain conditions, blood vessels may enlarge and their walls become thickened. To cite a single example, Nothnagel found that if the femoral artery of the rabbit is tied, the blood vessels, that come off immediately above the ligature, and which have already, through their subdivisions, connections in the muscles with other branches of the same femoral artery (that come off below the ligature), grow larger after a time. This he believes to be due, in the first instance, to the increased speed of the blood in the vessels, and thereby the bringing to these arteries of an increased food supply. Other writers have given different interpretations. Ziegler himself believes that several factors may be capable of bringing about the result. He thinks it improbable that the increase in the food supply can alone be the cause, and thinks it much more probable that the increased work that the vessels must perform while carrying more blood will account for the enlargement.

In connection with this discussion it may not be unprofitable to recall that in the regeneration of the lower animals we find simpler conditions in which proliferation of the cells takes place under circumstances where many of the factors suggested in the above discussion are absent. In the first place we find that new growth may occur without any increase in the nourishment that is brought to the organ. Regeneration takes place in the entire absence of food, except so far as it may be stored up in the tissues. Even in a planarian that is starving and decreasing in size, proliferation of new cells will take place if a part is removed. In many of the lower forms there may be proportionately even a much greater proliferation than in the regeneration and hypertrophy in the mammalian organs. It is true that proliferation may be more active if the tissues are well fed, but this does not show that the presence of food is a factor in the proliferation except so far as it keeps the proliferating cells in their best condition for growth. It is possible in many animals, more especially in some of the lower forms, to force them to grow rapidly by supplying them with a large amount of food, and conversely by decreasing the food to delay the growth. While this shows that the rate of growth is, within certain limits, a function of the amount of food, there may be also other factors that enter into the result, and in all cases there is an upper limit beyond which it is not possible to make the animal grow any larger.

That the presence of certain substances may bring about the enlargement of a part must be admitted as probable. It has been shown, for instance, that after the removal of certain lymphatic glands other glands may become larger. This appears to be due to the greater activity of the gland, brought about probably by the presence of an increased amount of some specific substance. In this instance the result can scarcely be due to a decrease in the physical resistance to growth or to an increase in the blood flow, except so far as this is brought about by the increased activity. It is, of course, possible, even if it cannot be positively shown in the case of the lymphatic glands, that a substance in the blood causes the hypertrophy in certain organs, while in others, as in the kidney, an increase in the blood flow may be also a factor in its hypertrophy.

The view held by several pathologists, that hypertrophy and regeneration may be caused by the removal of a physical resistance to growth, cannot be looked upon as a very probable hypothesis. The experiments in grafting of hydra and lumbriculus show that regeneration may still take place when the physical resistance has been reëstablished by grafting two pieces together. These results, which are more fully described in a later chapter, demonstrate that the growth is due to other influences.

A comparison with the lower animals shows that proliferation takes place when all but three of the factors considered in connection with hypertrophy and regeneration in the higher forms have been eliminated. These are, first, the action of substances that act either directly or as counteracting some substance already present, as Ziegler suggests; second, an innate tendency in the organism to complete itself; and, third, the use of the organ. It is impossible that the second factor enters into the problem of hypertrophy. In those cases in which regeneration takes place when a part of an organ is removed, as in the case of the liver, for example, the result may possibly also involve the second of the two factors, for the process is much like that of morphallaxis in the lower animals.

If it be granted that the growth in a hypertrophied organ is brought about by some substance that increases the function of that organ, can we suppose the phenomenon of regeneration to be due to similar factors? In other words, can we reduce both phenomena to the same principle? The case is complicated by two facts that may be illustrated by concrete examples. If a piece is cut from the middle of the body of lumbriculus new cells are produced at both ends of the piece. If we suppose the proliferation is brought about by the accumulation of certain substances in the piece, we must still invoke other factors to account for the differentiation of the proliferated material, since a head forms at one end and a tail at the other. All the hypothesis can do in itself is to account for a proliferation, not for the differentiation, and, both in the case of hypertrophy and in that of regeneration, it is the formation of new structures that we are chiefly concerned with, rather than the simple act of growth or of proliferation. If a piece of a hydra is cut off, the whole piece changes into the typical hydra form. Here there is no extensive process of proliferation, and the change is in the old part. It seems highly improbable that the production of substances in the piece could account for its change of form. These examples will suffice to show that in the process of regeneration it is very improbable that the change is brought about by special substances that may develop or be present in the part. We must suppose that during regeneration the formation of the typical form is not the result of a stimulus originating in a chemical substance acting upon the living material, but due to changes brought about directly in the living part itself. We must conclude, therefore, that despite the apparently close connection between the phenomena of hypertrophy of uninjured organs and of regeneration, they may often involve different factors.

If specific substances can bring about the hypertrophy of an organ, it is still not clear at present whether they do so by directly causing new growth, or whether their presence only stimulates the organ to greater activity and the activity of the organ is the cause of its growth. Since it must be supposed that in each organ a different specific substance brings about its activity and the consequent hypertrophy, it seems more probable that the result is due to the activity itself rather than to a stimulus from the substance. This view is further supported by the fact that in the case of the muscles and of the blood vessels the hypertrophy is directly connected with their use. The greater use brings about a larger supply of blood, but the blood is only different in amount and not in its quality. It must be confessed that it is difficult to see how the use of a part could make its growth increase, for by use the tissues break down; and we are not familiar with any other processes within the body that make for the building up of an organ in more than an inverse ratio to its breaking down. We are, however, familiar with phenomena of building up due to an increase in the food supply. It might appear from this to be more in accordance with what we find, to assume that the hypertrophy is solely due to an increase in the food supply; yet there are other facts known that show that an organ does not increase in size simply because it gets more blood, and that this occurs only when the organs have a greater functional activity. It is a safer conclusion, I think, at present to assume that both the activity of the organ and the increase in its supply of food acting together are factors in the result. On the other hand we are so much in the dark concerning the functioning and growth of organs that we can do little more, as the preceding pages show only too clearly, than speculate in the vaguest sort of way as to what changes take place; but since the processes seem to be within reach of experimental methods we can hope in the near future to learn more of how the processes of hypertrophy are brought about.

ATROPHY

It would not be profitable to enter into a general discussion of the many cases of absorption, or of atrophy of parts of the organism, but a few examples may be given that have a general bearing on the topics discussed in this chapter. The more noticeable cases arise through disuse of an organ, as shown, for example, in the decrease in size of the muscles of man when they are not used. Since this may take place in a single group of disused muscles, when no such change occurs in other muscles of the same individual that are in use, the most obvious explanation is that the decrease is due directly to disuse. Since the blood that goes to all the parts is the same, the diminution cannot be ascribed to any special substance in the blood. The flow of blood into the disused muscle is less than when the muscle is used, and it might be supposed that atrophy is directly caused by the lessened nourishment that the muscle receives. There is also the possibility that the decrease is brought about by the accumulation of certain substances in the disused muscle itself, but since, in general, the breaking down of the muscle is most active when it is used, it seems improbable that the result can be due directly to this cause, unless indeed it could be shown that the substances produced by a disused muscle are different from those in an active muscle.

Lack of food, as is known, may cause organs to decrease, the fat first disappearing, and then in succession in vertebrates, the blood, the muscles, the glands, the bones, and the brain. Certain poisons may also affect definite organs and bring about a decrease in size, as when the thymus and mammæ decrease from iodine poisoning, and certain extensor muscles after lead poisoning. Atrophy may also be brought about by pressure on a part, as when the feet or waist are compressed. In old age there may be a decrease in some of the organs, as in the bones, the testes and ovary, and even in the heart.

Degenerative changes appear even in the young stages of some animals, as when the tail of the tadpole is absorbed and the arms of the pluteus of the sea-urchin are absorbed by the rest of the embryo.

Especially interesting are the cases of absorption that take place when organs are transplanted to unusual situations in the body. Zahn transplanted a fœtal femur to the kidney, where it continued to grow but was later absorbed. Fischer transplanted the leg of a bird’s embryo to the comb of a cock, where it continued at first to grow, but after some months degenerated. The spleen, the kidney, and the testis have been transplanted, but they degenerate, and, in general, the larger the transplanted piece the more probable its degeneration. Small pieces of the skin have been transplanted from one individual to another, and it has been found that small pieces maintain themselves better than large pieces. Ribbert’s recent experiments in transplanting small pieces of different organs have been more successful than earlier experiments in which larger pieces were used. The first difficulty seems to be in establishing a blood supply to the new part, in order to nourish it. If the piece is quite small, it can absorb the substances, necessary to keep it alive, from the surrounding tissues, until the new blood supply has developed.

In the lower animals grafting experiments have been more successful, because the parts can remain alive for a longer time. It is important to find, however, that even in these cases, a part grafted upon an abnormal region of the body is usually absorbed. Rand shows that if the tentacles of hydra become displaced, as sometimes happens when a piece containing the old tentacles regenerates (Fig. 48, A-A³), the misplaced tentacles are absorbed; and I can confirm this result. In hydra, the hollow tentacles are in direct communication with the central digestive tract, and a displaced tentacle seems to be in as good a position as a normal one, as far as its nourishment is concerned, yet it becomes absorbed.

Rand also found, in other experiments, that when the anterior end of a hydra is grafted upon the wall of another hydra, the piece may maintain itself if it is large; but it is slowly shifted toward the base of the hydra to which it is grafted, and then the two separate in this region. If the graft is small, it may be entirely absorbed into the wall of the animal to which it is attached.

Marshall found that if the head of a hydra is partially split in two, each half-head completes itself (as Trembley had already shown). The body then begins slowly to separate into two parts, beginning at the angle between the two heads, until finally the two parts completely separate. King (1900) has repeated the experiment in a large number of cases with the same result. It seemed that the division might be brought about by the weight of the halves causing the gradual separation of the body, but King has shown that this is not the case, for, when a double form remained hanging with its head down, it still divided into two parts (Fig. 47, A). In this case, the weight of the two heads would cause the parts to come together rather than to separate, if gravity had any influence of the sort suggested. Marshall and King have also shown that if the posterior end of a hydra is split in two, the two parts do not continue to separate, but one of the two, if the pieces have been split some distance forward, may become constricted from the other, and, producing new tentacles at its apical end, become a new individual.